The most straight forward way to put it from a technical point of view is the cancelation that occurs at a single point in space (like a microphone) when identical frequencies arrive at different times. Each frequency where this occurs creates another cancelation tooth in the comb that is visible in the measurement.
It can be caused by playing a pair of speakers and positioning the microphone even just inches off centre or by the reflected sound arriving at a different time from the fundamental. The fact is that we are listening to a pile of comb filtering all the time when we play our audio systems and there is lots of empirical testing that has been done to show this creates a favourable listening experience.
I would say there are three camps of opinion on comb filtering:
There is the camp that believes all comb filtering is bad. Unfortunately the only way to accomplish zero comb filtering would be to listen in mono in an anechoic chamber. Having done this myself I can attest that it is not a great listening experience.
Then there is a camp that tries to separate the comb filtering into two or three groups and rate each group separately as to its effect on the listening experience. The most benign of these groups would be the comb filtering caused by the reflected sound. Next would be the comb filtering caused by using a stereo pair of speakers (or more) and not sitting exactly in the centre. Lastly, and this camp believes this type is bad, is the comb filtering that occurs when two identical drivers (usually tweeters but that depends on the physical spacing) in one speaker are used. I have seen some highly technical explanations of this difference but none that are backed up with any empirical scientific testing.
Then there is the camp that we are a member of which embraces all forms of comb filtering as either favourable or at worst irrelevant. And we have conducted empirical scientific testing to back this up.
Last edited by Ian; 03/24/14 07:49 PM.
